reliability

• reliability 


  reliability is a measure of consistency, if the results are not consistent then the measure is not reliable, can be assessed in two ways: test-retest and inter-observer
• test-retest reliability 


  when the same person or group are asked to undertake the research measure on a different occasion

  after the measure has been completed on two separate occasions, the two scores are then correlated, if the co-efficient is over +0.8 the research measure has strong reliability
• limitation of test-retest reliability

  assessing reliability using test-retest method is that demand characteristics can occur as they could be recalling what they put down previously

  it can also be time consuming as the second test could take place over a year later
• inter-observer reliability 


  when 2 or more observers are recording behaviour in a consistent way, behaviour categories are used in observations which are subjective, so when the separate observers correlate their observations, there should be a similarity
• example of inter-observer reliability

  Ainsworth's strange situation (1978) found 98% agreement across the observers on the behavioural categories observed, this makes the findings more meaningful
• improving reliability - interviews

  can be improved by reducing the researcher bias by having the same interviewer for all interviews, switching to structured interviews
• improving reliability - experiments

  can be improved by standardising procedures so there is more control extraneous variables so they do not become cofounding variables
• improving reliability - observations


  can be improved by operationalising behavioural categories (making them measurable) to reduce subjectivity and increase objectivity